How Climate and Deep Tech Startups Get Funded

A climate company raises a strong Series B, proves its technology at pilot scale, and hits a wall that software companies never encounter.

The next step is not a bigger office. It is a first commercial plant costing considerably more than the entire venture capital raised to date. Venture investors will not fund it — it is an asset, not a company, and it would consume a fund. Project financiers will not fund it either — the technology has no operating history, so the risk is not one they price.

This is the first-of-a-kind problem, and it is the defining feature of financing anything physical. This guide covers the full capital stack, how offtake agreements function as financing, the catalytic capital that exists specifically to bridge the gap, what investors actually underwrite, and how the milestones differ from software.

1. Why Venture Capital Alone Cannot Do This

Venture funds companies with a small number of enormous outcomes covering many failures. The model assumes capital efficiency and fast iteration, neither of which describes building a plant.

Project finance is the opposite: lending against a specific asset with predictable cash flows, secured on the asset itself, priced for low returns and low risk. It requires proven technology, a creditworthy buyer, and a long operating record.

A first commercial facility sits between the two. Too capital-intensive and too asset-like for venture; too technologically novel for project finance. Founders who do not recognise this early plan a Series C that no venture fund will write.

The practical response is a capital stack that layers different money against different risks.

2. The Capital Stack

Early venture equity

Funds the science, the team, the pilot. Conventional venture terms from deep tech and climate specialists, though usually with longer expected timelines and more tolerance for hardware.

Non-dilutive government funding

Substantial in this sector and consistently under-used.

• SBIR and STTR, particularly through the Department of Energy and NSF
• Advanced research agency programmes funding high-risk, high-reward energy and industrial technology
• Demonstration and deployment programmes, which fund pilot and first commercial facilities at a scale grants rarely reach
• Government loan programmes providing debt for projects commercial lenders will not yet finance — slow, demanding, and transformational when they land
• State and regional programmes, frequently tied to siting a facility locally, which can be negotiated alongside land and utility arrangements
• Allied and international programmes for companies operating across borders

Strategic and corporate capital

Industrial companies, energy majors, utilities, chemical and materials producers invest for access to technology. They bring engineering capability, site access, offtake and eventual acquisition — and the same cautions apply as elsewhere in corporate venture, particularly on exclusivity and rights over your technology.

Catalytic and concessionary capital

Money that accepts below-market returns to unlock other capital — from foundations, philanthropies, development finance institutions and climate-specific vehicles. It typically takes a first-loss position, which is exactly what makes conventional investors comfortable enough to participate. This is the least understood and most useful part of the stack for a first plant.

Project debt and infrastructure equity

Once technology risk is retired, projects finance conventionally — and this is where the very large money is. Infrastructure funds and project lenders will fund plant two, five and twenty on terms that make no sense for plant one.

Equipment financing

Specific machinery with a resale market can be financed as an asset rather than funded from equity, which is dramatically cheaper capital. Covered in our guide to private credit and asset-backed lending.

Tax-driven structures

Tax credit and incentive regimes have historically been central to financing energy and industrial projects, including structures that let credits be monetised by parties other than the project owner. This area has been repeatedly amended and remains in flux — several incentives have been modified or phased down in recent legislation. Do not build a financing plan on a credit without current specialist tax advice on its status and duration.

3. Offtake Agreements Are Financing

The single highest-leverage thing a climate or industrial company can obtain is not a term sheet. It is a signed offtake agreement — a creditworthy buyer contractually committed to purchase your output at a defined price for a defined period.

An offtake converts an unfinanceable project into a financeable one, because it gives lenders exactly what they need: predictable revenue from a counterparty they can assess. A project with a fifteen-year contract from an investment-grade buyer is a fundamentally different proposition from the same plant with a market-price assumption.

Forms this takes:

• Power purchase agreements for energy
• Product offtake for materials, fuels and chemicals
• Advance purchase commitments, where a buyer pays ahead of delivery — genuinely non-dilutive working capital
• Carbon removal prepurchase, where corporate buyers pay in advance for future delivery, which has become a meaningful funding source for early removal companies
• Volume commitments without fixed price, weaker but still useful

The practical advice: pursue offtake with the same seriousness as fundraising. A commercial team that lands a credible anchor buyer has done more for your capital position than an additional venture round would, and at zero dilution.

4. What Investors Actually Underwrite

• The cost-down curve. Not today's cost — the engineered path to a competitive price at scale, with the specific levers named: throughput, yield, materials, energy intensity, capital cost per unit of capacity. Vague appeals to learning curves are not persuasive.
• The green premium, honestly stated. How much more does your product cost than the incumbent, who is willing to pay it, and what closes the gap — scale, policy, or a buyer's own commitments?
• Scale-up risk. Chemistry and physics behave differently at commercial scale. Investors want evidence that engineering risk has been identified rather than assumed away.
• Permitting and siting. Frequently the longest pole. Interconnection queues, environmental permits, local approvals — timelines measured in years and largely outside your control.
• Supply chain for critical inputs, and concentration in any single source or jurisdiction.
• Policy dependency. A business that only works with a subsidy is a bet on the subsidy persisting. Investors have been burned by this and now test it directly. Show what the business looks like without it.
• The team's build experience. Scientists who have never commissioned a plant are a real risk. Investors look for people who have delivered physical projects.

5. The Milestones That Actually Move Valuation

Software companies raise on growth metrics. Climate and deep tech companies raise on de-risking events, and knowing which ones matter changes what you spend money on between rounds.

Technology readiness, demonstrated not asserted. The ladder from bench to pilot to demonstration to commercial scale is what investors actually track, and each step is a genuine valuation event. What does not move anything is spending eighteen months improving a bench-scale result. Investors have learned that laboratory performance frequently does not survive scale-up, so the money follows the step change, not the incremental improvement.

Independent verification. A third party confirming your performance data is worth substantially more than your own measurements, and companies consistently underinvest here. For a technology whose whole case rests on an efficiency or a yield figure, an independent test report is one of the cheapest credibility purchases available.

The first offtake. Covered above, and worth restating as a valuation event rather than a commercial one. A signed contract from a creditworthy buyer changes what kind of capital you can access, which changes your options more than any revenue number.

A site with permits in hand. Land, interconnection, and the environmental approvals for a specific location. This is unglamorous, takes years, and is one of the strongest signals a company is real — precisely because it cannot be accelerated with money or enthusiasm.

A named engineering partner. An established engineering, procurement and construction firm willing to work with you — and eventually to wrap the build — tells investors that people who build things for a living have looked at your design and are prepared to stand behind it.

The first plant operating. The single largest step, because it converts you from a venture-risk company into a project-developer company with an operating record. Everything after this is a different financing conversation.

The practical implication: plan the capital between rounds around reaching the next de-risking event, not around a runway number. A company that raises eighteen months of runway and arrives at month eighteen having improved a bench result is in a much weaker position than one that arrives having permitted a site.

6. Sequencing That Works

1. Fund the science non-dilutively as far as it will go. Grants for early technical work preserve equity for the expensive part.
2. Prove unit economics at pilot scale, with instrumented, credible data.
3. Land an anchor offtake before you need project capital. This is the step most companies leave too late.
4. Assemble a blended stack for the first plant — grant, catalytic first-loss, strategic equity, government debt. Expect this to take longer than a venture round and to involve parties who have never worked together.
5. Operate, and generate the record that makes plant two conventionally financeable.
6. Move to project finance for replication, where the capital is deep and comparatively cheap.

The mental shift that matters: after the first plant, you are no longer raising venture capital. You are developing projects, and the skills, counterparties and documents are entirely different.

7. Common Failure Modes

• Planning a venture round for a plant. No venture fund will write it, and discovering that with nine months of runway is fatal.
• Treating permitting as an afterthought. It is frequently the critical path, and it cannot be accelerated with money.
• Assuming a policy incentive is permanent. Incentive regimes change with governments, and recent years have proved it repeatedly.
• No offtake before seeking project capital. Lenders need contracted revenue; hope is not a counterparty.
• Underestimating working capital. Long lead times on equipment mean paying suppliers well before revenue.
• Over-diluting early. Capital-intensive companies raise many rounds. Founders who take heavy dilution at seed have very little left by first plant — which makes the preference stack a live problem long before an exit.

Frequently Asked Questions

Is impact capital different from ordinary venture?

It varies enormously. Some impact funds seek fully market-rate returns with an additional screen; others explicitly accept lower returns for measurable outcomes. The second group is genuinely valuable in a first-plant stack because it can take positions others will not. Establish which you are talking to early — the conversation and the terms differ.

How long until an exit?

Longer than software, typically. Acquisition by an industrial or energy incumbent is the most common route; public listing happens for companies at genuine scale. Some companies become project developers with recurring economics, which is a different and legitimate outcome that suits infrastructure investors rather than venture ones.

Should we build or license our technology?

Licensing to established manufacturers avoids the capital requirement entirely and can reach scale faster, at the cost of margin and control. Building captures the economics and requires the hardest financing. A hybrid — build the first plant to prove it, licence thereafter — is a common and often sensible answer.

Do carbon credits count as revenue?

Investors scrutinise this closely. Prepurchase agreements from credible corporate buyers are real, contracted revenue. Speculative credit sales into a voluntary market with uncertain pricing and evolving standards are discounted heavily. Show which you have.

Can we use venture debt?

Sometimes, though hardware companies without recurring revenue are a difficult credit and warrant coverage tends to be higher. Equipment financing against specific machinery is usually the better instrument — our guide to venture debt covers when each applies.

How much should we raise at each stage?

Enough to reach the next de-risking event with margin, which for hardware means more than founders instinctively plan for. The specific trap is raising to a runway figure rather than to a milestone: equipment lead times, commissioning delays and permitting slippage are normal in this sector, and a company that budgeted no contingency arrives at the milestone three months late with no cash. Add meaningful buffer to any physical timeline, and be explicit with investors about why — experienced hardware investors respect it and inexperienced ones need to hear it.

Do we need a project development team, and when?

Earlier than most companies hire one. Site selection, permitting, interconnection and utility negotiation are specialist skills with long lead times, and they are not something a technical founder can pick up alongside running the company. The usual pattern is a first development hire around the time the pilot is working, well before there is a plant to build — because the work they do determines whether a plant is possible at all two years later.

How do investors think about the first plant failing?

Realistically, and it is worth discussing openly rather than avoiding. First-of-a-kind facilities frequently underperform initially — commissioning takes longer, throughput ramps slower, and yields start below design. Sophisticated investors expect this and budget for it. What damages a company is presenting a plan with no allowance for it and then requiring emergency capital when reality arrives. Show the ramp curve you actually expect, including the bad months.

Does an industrial partner limit our exit options?

It can, and the terms matter more than the relationship. Rights of first refusal, exclusivity in a territory or application, and restrictions on working with a partner's competitors all narrow the eventual buyer pool — sometimes to one. That is not automatically wrong; a strategic partner who becomes the acquirer is a legitimate and common outcome in this sector. It is wrong if it happens by accident because nobody read the clause. Negotiate exit-affecting provisions with the same care as the economics.

The Bottom Line

The hard part of climate and deep tech is not the science and not the seed round. It is the first commercial plant, which fits neither venture nor project finance and must be assembled from several kinds of capital at once.

Fund the science with grants, land an offtake before you need project money, plan capital around de-risking milestones rather than runway, and be honest about the cost-down curve and the policy dependency. Investors in this sector have been burned by optimism and reward precision.

Global Capital Network connects climate and deep tech founders with venture, strategic and infrastructure investors across our network and events. Get in touch.

This article is general information, not financial, legal or tax advice. Incentive regimes in this sector change frequently. Verify current programme status with specialist advisers.